Absolute measures of association

Absolute measures of association

While measures of incidence and prevalence may provide valuable information about the characteristics of a population, we will often be interested in the association of an exposure with a disease, event, or other outcome to make inferences about potential causal effects. In general, this requires comparing exposed and unexposed individuals to identify an association between an exposure and a given outcome.⁴

The risk difference provides an absolute measure of the difference in risk between an exposed and unexposed group. Risk difference can be calculated as: RD=R₁-R₀, where RD is the risk difference and R₁ and R₂ represent the risk of the outcome in the exposed and unexposed groups, respectively.⁴ When the groups compared have a similar risk of the outcome aside from the exposure of interest, this will provide a valid estimate of the risk difference.⁴ In practice, epidemiologic studies will typically need to use either randomization or sophisticated observational study designs to ensure that exposed and unexposed groups in a study are comparable and to adjust appropriately for differences.

A study of the association of the relative age of children in a school grade with diagnosis of attention deficit hyperactivity disorder (ADHD) included estimates of risk difference.¹² Due to the cut-off date for enrolment of children in school in British Columbia, children born in December would typically be almost a year younger than children born in January in the same grade. Among boys, 7.4% of those born in December received a diagnosis of ADHD, compared to 5.7% of those born in January. Among girls, 2.7% of those born in December and 1.6% of those born in January received a diagnosis of ADHD. The association between relative age of children in a school grade and diagnosis of ADHD could be shown by calculating the risk difference of 1.7% for boys and 1.1% for girls. The authors suggested that the relative age effect in diagnosis of ADHD might reflect over-diagnosis of younger children due to their immaturity.¹²